1
Enhanced Public Health Reporting
Using an HIE Network
Session 143, February 13, 2019
Brian E. Dixon, PhD, FACMI, FHIMSS, Director of Public Health Informatics,
Regenstrief Institute, Inc. and Associate Professor,
Indiana University Richard M. Fairbanks School of Public Health
2
Brian E. Dixon, PhD, FACMI, FHIMSS
Twitter: @dpugrad01
Has no real or apparent conflicts of interest to report.
NOTE:
Dr. Dixon is part-time VA employee. Comments are personal and
should not be attributed to the Department of Veterans Affairs or the
Federal Government.
Conflict of Interest
3
• Case Reporting for Notifiable Disease
– Historical perspective
– Significance for clinical and public health practice
• Controlled Before-and-after Trial of HIE-based Intervention
– Indiana Health Information Exchange
– Study Design and Methods
– Results
– Discussion
• Conclusions
Agenda
4
• Describe the barriers to timely, complete reporting of notifiable
diseases to public health authorities
• Discuss the policies and requirements for reporting information to
public health agencies
• Define the concept of electronic case reporting in support of public
health
• Explain how a health information exchange network can facilitate
electronic case reporting
Learning Objectives
5
• Computer-based clinical decision support (CDS) can be defined
as the use of the computer to bring relevant knowledge to bear on
the health care and well being of a patient.
– Greenes, 2007
Clinical Decision Support
Friedman, JAMIA, 2008
6
• Public health decision support (PHDS) can be
defined as the use of the computer to bring
relevant knowledge to bear on the health and
well-being of a population.
– Dixon, Gamache & Grannis, 2013
– doi.org/10.1136/amiajnl-2012-001514
• Examples:
– Vaccine forecasting report
– Suggestion for ordering stool culture
Public Health Decision Support
7
Public Health Case Reporting
8
• “for which regular, frequent, and timely information regarding
individual cases is considered necessary for the prevention and
control of the disease” McNabb, et al., 2008
• Examples of notifiable diseases
– HIV / AIDS
– Sexually transmitted infections (e.g., Chlamydia)
– Enteric diseases, including E. coli, Salmonella
– Lead poisoning
– Zika virus
– Lyme disease
Notifiable Diseases
9
• Most states require reporting of notifiable diseases
– State law varies with respect to disease, requirements
• Notifiable disease reporting often uses manual, spontaneous
reporting processes
– Paper, Phone, Fax
– Relies on providers, labs to Identify and Report
• Varied workflow at health department based on disease
– Routine (e.g., chlamydia)
– Intense (e.g., HIV)
– Dixon et al. 2014, 10.5210/ojphi.v5i3.4939
Notifiable Disease Case Reporting
10
Traditional Case Reporting Workflow
© Regenstrief Institute
11
• Between 9% and 99% cases reported (high variance)
– Most diseases less than 20% cases
– Doyle et al., 2012, Am J Epidemiol
• Why care about disease reporting to public health?
– Accurate reporting of disease burden (epidemiology)
– Timely control and response
– Cost of care for rising incidence ($$$)
– Antibiotic resistance
Problem: Provider Underreporting
12
• In pre-intervention survey, 60.7% of clinic staff (N=29) said they
had previously reported to PH
• I need to report that to public health?
– Lack of awareness (28%)
• I don’t know to whom or how to report that…
– Lack of understanding of process (21%)
• No one’s fined me for not reporting that…
– Lack of sufficient rewards/penalties
PH Reporting: Provider’s View
13
• Leverage health information technology (IT)
components available in our ecosystem
• Implement a solution that minimizes burden on
clinics while maximizes yield for public health
organizations
• Utilizes available standards in support of
interoperability
How can we improve provider
reporting rates?
14
Official State Case Report Form
patient
Information
Name
Address
Phone#
DOB
Gender
Race/ethnicity
lab
Information
Etiologic agent
Test name
Test date
Treatment initiation date
Treatment (drugs)
provider
Information
Physician name
Physician address
Phone#
Reported by
Report date
15
Data Management
Data
Repositor
y
Health
Information
Exchange
Network
Applications
Payers
Physician Office
Ambulatory Centers
Public Health
Hospital
Labs
Outpatient RX
Data Access & Use
Hospitals
Physicians
Labs
Public
Health
Payer
• Results delivery
• Secure document transfer
• Shared EMR
• Credentialing
• Eligibility checking
• Results delivery
• Secure document transfer
• Shared EMR
• CPOE
• Credentialing
• Eligibility checking
• Results delivery
• Surveillance
• Reportable conditions
• Results delivery
• Secure document transfer
• De-identified, longitudinal
clinical data
Researchers
The Indiana Network for Patient Care
16
Quick Stats on the INPC
• 117 hospitals, representing 38 health systems
• Over 16,000 practices with over 45,000 providers
• Over 14 million patients
• Nearing 12 billion pieces of clinical data
– Doubled in the past 2 years!
17
The Notifiable Condition Detector
Fidahussein M, Friedlin J, Grannis S. Practical Challenges in the Secondary Use of Real-World
Data: The Notifiable Condition Detector. AMIA Annu Symp Proc. 2011:402-8.
18
• Existing HIE communication pathways
– Electronic laboratory reporting (ELR)
• Automated case detection
– Identification of cases that should be reported to PH
– Classification of disease using LOINC / SNOMED CT
• Clinical messaging (aka DOCS4DOCS @IHIE)
– Getting information to its recipient in a way that is
integrated into workflow
Leveraging Robust Infrastructure
19
Enhanced Case Reporting Workflow
© Regenstrief Institute
20
Pre-Populated Notifiable Report
21
• Controlled Before-and-After Study
– Intervention clinics (N=7) were not randomized, but
there were concurrent controls (N=312)
– All clinics were connected to INPC via D4D
• Timeframe: 2013-2016; Setting: Indianapolis, Indiana
• Difference-in-difference analysis to detect ∆
– Focus is ∆ between intervention and control sites
– Binomial GLM with logit link function and NLEstimate
macro
Study Design and Methodology
22
• Source of Data: Case files from the Marion County Public
Health Department
– All cases for 7 representative diseases: CT, GC, HBV,
HCV, Histoplasmosis, Salmonella, Syphilis
– Case records include lab, HIE, and provider reports
• A report is a fax, paper report, or e-report
– We looked at reports as well as the fields within the
report, such as patient name, address, lab test, etc.
• Goal: Comprehensive review of all reports for each case as
well as the information in each report
Data and Sources
23
• Primary Outcome
– Provider Reporting Rate: the proportion of cases
where there is at least one report from a provider
(clinic or hospital)
– Remember that the lab can also submit reports
• Secondary Outcomes
– Completeness of key fields used by disease
investigators: the proportion of non-null values
received by MCPHD
– Timeliness of reports: Difference in # days between
lab result and when report submitted to MCPHD
Outcome Measures
24
Results of Evaluation
25
Provider Reporting Rates
12.40%
20.20%
10%
50%
-10%
10%
30%
50%
70%
Control Clinics Intervention Clinics
Before
After
p <.001
26
Reporting Rates Over Time
27
Provider Reporting Rates (Chlamydia)
28.80%
56.90%
21.70%
66.40%
0%
10%
20%
30%
40%
50%
60%
70%
Control Clinics Intervention Clinics
p <.001
28
Provider Reporting Rates (Gonorrhea)
27.50%
55.60%
20.60%
58.30%
0%
10%
20%
30%
40%
50%
60%
70%
Control Clinics Intervention Clinics
29
Provider Reporting Rates (Hepatitis C)
6.40%
6.50%
2.00%
7.30%
0%
10%
20%
30%
40%
50%
60%
70%
Control Clinics Intervention Clinics
Before After
30
• 4 of 15 Fields Significantly Improved (p<0.001)
– Physician First Name, Last Name
– Physician Address, Zip Code
• 9 of 11 Remaining Fields Improved**
– Patient Information, Lab Test Performed
– Completeness from control clinics also improved
• Patient First and Last Name Remained 100%
Completeness of Data in Reports
31
Timeliness of Provider Reporting
11.25
10.13
7.96
9.67
0
2
4
6
8
10
12
Control Clinic Intervention Clinics
Days
Before After
32
• Alerting clinics to new cases of notifiable disease is feasible
and effective at improving reporting rates
– Clinics responded to alerts with submissions to the
LHD and provided more complete reports*
• The intervention effects were not uniform
– Timeliness of reporting did NOT change
– Chlamydia benefited the most
– Other diseases did not improve significantly**
Trial Conclusions
33
• Leverage existing standards and pathways where possible
– Use of LOINC and SNOMED CT in ELR messages
– Utilize eCR C-CDAs and FHIR APIs where they exist
• Public health services part of an HIE network are not always
revenue generating
– Policy or other drivers might be necessary to drive adoption
• Solutions should fit into clinic workflow
– Current solutions for “outside” information not optimal
– EHR systems should assume coordination with external
entities such as public health departments
Lessons and Discussion
34
• Electronic case reporting (eCR) is a public health option specified
in Stage 3 meaningful use
– Also MIPS Public Health Reporting criterion
• If we can alert providers to cases that should be reported and
enabled electronic submission of reports, we should see reporting
rates increase across diseases
– PH Decision Support combined with MU functions
• Do not focus solely on MDs / physicians
– Clinic “reporters” are nurses, MAs, others
– Revere et al., 2017. doi.org/10.1186/s12889-017-4156-4
Implications of Trial
36
• Our Study Team
– Shaun Grannis, MD (IUSM and Regenstrief)
– Zuoyi Zhang, PhD (Regenstrief)
– Joe Gibson, PhD (Marion Co. Public Health Dept.)
– Debra Revere and Becky Hills (U. Washington)
– Patrick Lai, MPH (SOIC) and Uzay Kirbiyik (FSPH)
– Abby Church, PMP (Regenstrief Institute)
• The work presented was supported by grants from
– AHRQ (R01HS020209) and
– RWJF (71596) part of the PHSSR Portfolio
Acknowledgements
37
Public Health Informatics Program
@Regenstrief Institute
Support and Improve the Business of Public Health
• Automating reporting of cases (ELR, ECR) to PH agencies
• Leveraging EHR data for chronic disease prevalence
Assess and Improve the Health of Populations
• Improving vaccination rates and population immunity
• Reduce the proportion of children who are overweight
Educate and Train the Next Generation
• Provide high quality informatics education to MPH, MD, etc.
• Train the future leaders of public health informatics
38
Brian E. Dixon, MPA, PhD, FACMI, FHIMSS
Associate Professor, IU Fairbanks School of Public Health;
Director of Public Health Informatics, Regenstrief Institute;
Health Research Scientist, Department of Veterans Affairs
http://bit.ly/bedixon
Twitter: @dpugrad01
Email: bedixon@regenstrief.org
Questions and Discussion
